255
Brain Research, 592 (1992) 255-260 © 1992 Elsevier Science Publishers B.V. All rights reserved ~06-8993/92/$05.00
BRES 18155
CNQX increases spontaneous inhibitory input to CA3 pyramidal neurones in neonatal rat hippocampal slices Chris J. M c B a i n , J a m e s V. E a t o n , T r a c y B r o w n * a n d R a y m o n d D i n g l e d i n e * Department of Pharmacology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599 (USA) (Accepted 19 May 1992)
Key words: CNQX; Hippocampal slice; Inhibitory postsynaptic current; Whole-cell recording; Interneuron
Whole-cell recordings were made from immature CA3 pyramidal neurones in the rat hippocampal slice. The addition of the giutamate receptor antagonist, CNQX, caused a robust increase in the frequency of spontaneous inhibitory Dost-synaptic currents (IPSC) concomitant with the expected reduction of excitatory drive to these neurones. This effect of CNQX was not shared by structurally related quinoxalinediones or ~nurcnic acid, which arc a l ~ antagonists of non-r,IMDA glutamate receptors. This effect of CNQX was abolished by tetrodotoxin suggesting that an increase in interneurone spiking was responsible for the IPSCs. Recordings from stratum radiatum interneurones of CA3 confirmed this suggestion, since some interneurones were directly depolarized by CNQX. The excitation by CNQX of a small population of stratum radiatum interneurones of CA3 complicates interpretation of experiments designed to assess the consequences of blocking excitatory transmission with this
drug.
INTRODUCTION
Certain quinoxalinediones substituted at the 6- and 7 position are competitive antagonists at non-NMDA glutamate receptors 4''~'~'14'1¢' (for review see ref. 30). The ability of the nitro-substituted quinoxalinediones, CNQX and DNQX, to block polysynaptic excitatory pathways in the rat hippocampus I'z'4'24 has made them useful tools in the study o! other receptor systems involved in synoptic transmission, particularly the NMDA and GABA receptor complexes (e.g. see ref. 1). These antagonists also dampen a variety of epileptiform tt't6-ta'23 and excitotoxic '°'2s activity in the brain. Some quinoxalinediones also block the NMDA receptor complex 3 by a competitive antagonism of the glycine co-agonist site ~'~3't6'29. Recently, Jarolimek and Misgeld t5 showed that CNQX at high concentrations (10-50 /~M) also antagonized GABA A receptors on cultured hypothalamic and mid-brain neurones. Here we describe a novel effect of CNQX not shared by other structurally related quinoxalinediones. Using the whole-cell patch clamp technique in slices of
neonatal rat hippocampi, we have observed that, at concentrations sufficient to antagonize excitatory input, CNQX causes a robust increase in frequency of spontaneous inhibitory post-synoptic currents (IPSCs) received by the CA3 pyramidal neurone population. Recordings from stratum radiatum interneurones suggest that depolarization of a small population of local interneurones is responsible for this observation. MATERIALS AND METHODS Rat hippocampal slices (300 /~m thick) were prepared from 8-15-day-old Sprague-Dawley rats, although a few cells were recorded from earlier ages, as described previously I'~. Slices were held in a submerged chamber and perfused at a rate of ~ I-3 ml/min with the following medium (in raM): NaCI 130', NaHCO.~ 24; KCI ~.5; NaH:PO4 1,25; CaCIz.2HzO 1.5; MgSO4.7tt20 1.5; glucose 10; saturated with 95% 0 : - 5 % CO: (pH 7.4, 307 mOsm) at 22-25°C. All drugs were bath applied in known concentrations by direct addition to the perfusate via a three way tap. The dead time of the perfusion system was - t rain. Tight-seal ( > I G ~ ) whole-cell recordings 7'12 were obtained from the cell bodies of 55 neurones in stratum pyramidalc of CA3, and 12 interneurones located in stratum radiatum of CA3. Patch ~lectrodes were fabricated from borosilicate glass and were not fire polished. They had resistances of 3-7 M£~ when filled with (in mM); caesium methanesulphonate 140; HEPES 10; MgCI2 2 (buffered to pH 7.3
Correspondence: C. McBain, National Institutes of Health, Bldg. 36 room 2A21, Bethesda, MD 20892, USA. *
Present address: Department of Pharmacology, Emery University School of Medicine, Atlanta, GA 30322, LISA.
256 with CsOtl, 275 mOsm). Breakthrough to the whole-cell mode was perfi+rmed under current chimp in order to provide an initial evaluation of the passive and active properties of the neurone membrane. Cells possessing an initial resting potential more negative than - 4 5 mV (pyramidal neurones - 57_+ 0.9 mV, . = 55: interneurones - 52 ± 2.5 mV. n = I I ) and overshuoting action potentials were accepted, Diabsis of the neurone by Cs + usually resulted in the cell stabilizing at a less negative potential. Neurones that fired action potentials on the anode break were discarded. Unless stated otherwise cells were voltage clamped at - 4 0 or - 5 0 mV to permit discrimination between excilato~' and inhibitory synaptic currents. Input resistances were 285 _+23 M.f~ (n = 55 pyramidal cells) and 411 ± 42 M.Q (n = I I interneurones) at this holding potential. All currents were recorded with an Axopatch-ID amplifier (Axon Instruments): re~ords were filtered at i - 3 kHz ( - 3 dB)and digitized ;it 3-10 kHz on a 80386 computer or stored on FM tape (5 kHz b:mdwidth) fi~r off-line analysis. Individual EPSCs and IPSCs were subjectively identified and accepted using the following criteria: (1) the mean current associated with the event was significantly different from that of an equivalent length of pre-cvent baseline; (2) currents were aysmmetrical with a monotonic fast rise and approximately exponential decay. Measurements u[ the decay time constant of synaptic currents r:prescnt the mean of at least 15 individual events from each aeurone and were fit by one exponential component (simplex algorithm, least squares criteria), Amplitude histograms were constructed from individually measured IPSC amplitudes and hinned in 2 pA intervals. The series resistance was determined by balancing the "bridge' in current chimp, and lypically ranged from 10-3() M.f2, The amplitude of the inhibitory synaptic currents was never in excess of 80 pA, thus a series resisl.'lnce of 25 M.f~ would rcsuh in a series resistance error of no greater thai} 2 mV. Membrane potentials were not corrected for these errors, The steady-state current-voltage (I-V) relation of hath-applied CNQX was obtained by measuring the current ;It the end of 25()-I,1}()() ms voltage st~ps applied in = 10 mV increments at 10 s inlcrvals, A holding polential o f ~1=5|l n'lV W,'IS selected for Hlese stltdh.~s to inacliV;|le most voltagedependent currents. I=V rehithms were obt;fined prior to, durin~ :tddilion and fi)llowing rentoval of k;finate, The mean of Ihc control and recovery was Ihun sul'qra¢led from that in CNQX, All data nr¢ represented ,'is the: mean i S.E,M, Where approprt. ate. a Stndcnt's t.test wits pcrfornlcd, I)rugs used wcrc tetrodotoxin ('rl'X, Si~,nla), hicttcuIline nl~tho. I','omide (Sigma }. h.cyarlo.7.nitro.quinoxalinc.2,3.dione (('NQX} and h.7.dinit|oquitloxuline.2,3.dione (DNQX, Toeris Neuramin), kynurenic a~.'id, (Signla), and 2,.~.dihydroxy.6-nilro.7.sulfin'noyl. benzo(F) quinoxalinc (NBQX, generous gift from T, I lonore):
RESULTS CN(.)X increases spontaneous IPSC fi'equency hz CA3 pyramids Pyramidal ncuroncs were voltage clamped at - 40 to - 5 0 mV, which permitted the separation of inhibitory (outward at this holding potential) from excitatory (inward) synaptic currents (Fig, 1). The pyramidal ncutones received spontaneous IPSCs at a frequency of 0.3-17.1 Hz (mean 3.4:t:0,6 Hz, . = 3 4 ; Table I). These IPSCs ranged in amplitude from 5 to 80 pA. The falling phase o1" the IPSC could be fit by a single exponential with a time constant of ~ 20 ms (Table !), EPSCs were not studied further and have been described in detail eisewheret". In 3(} of 34 neurones addition of CNQX (3 pM) to the bathing medium virtually abolished EPSC activity while increasing the frequency of IPSC activity (Figs. I
>,70t
A. Control
60f 5of
.
. .
=o',+of o 20f
cn 1Of
E.itlii
Q- | ~~ __ 00 10 20 30 40 50 - -
:=,.,70 B. 3.aM CNQX
sot g'+ot 3of
u2°t
~nlo+
0
I l L ..lit.
~
_
10 20 3'3 40 50
"
E+'O~lS°sI°l
~70,C. 3#M CNOX+ I/zM Bicuculline
,,.-:3ot o 2o1
u~ 10,1.
OOL 10 20 3 0 40 IPSC Amplitude (pA)
50
Fig, 1, CNQX increases the frequency of spontaneous IPSCs and blocks spor{aneous EPSCs. A representative recording from a stratum pyramidale CA3b neurone. IPSC amplitude histograms (bin size = 2 p a l are shown on the left. The histograms were generated from 59 s of data. In the current traces shown on the right, upward deflections represent IPSC:s and downward deflections represent EPSCs, Those evenls marked ' + ' were accepted as IPSCs, CNQX induced :k suhst:mlial increase in IPSC frequency while having no noticeable effect on IPSC amplitude or decay time constant, A: control (ACSF only); m~an amplitude ,= 17 pA, ntean decay constant ., 23 ms, IPSC frequency + 3,8 |lz, B: 3 / ~ M CNQX; IPSC amplitude ,~ 21 pA, decay lint~ constant +. 24 ms, IPSC frequency =9.2 llz, C: most IPSCs arc clitniflated by the addition of bicuculline (5 # M)in the presence of C'NQX,
an0 2). This effect of CNQX was not accompanied by a~y change in the pyramidal neurone holding current or input resistance, and CNQX was without significant effect on the r~lean amplitude or decay time constant of IPSCs (Table !). This effect was readily reversible and could be reproduced by re-application of CNQX (Fig. 2). The synaptic currents in CNQX could be blocked by the addition of the specific GABA^ receptor antagonist, bicuculline (5/zM; Fig. IC), or tetrodotoxin (Fig. 2El, suggesting that the IPSCs were caused by ongoing interncurone activity. The magnitude of the increase in IPSC frequency by CNQX varied greatly from cell to cell, from less than 2-fold to about 25-fold. A plot of IPSC frequency in the absence of CNQX vs. that in the presence for all cells tested (Fig. 3) indicates that this wide range observed is not due to a 'ceiling effect'. Rather, CNQX appeared to increase IPSC frequency by 2 - 3 / s regardless of the initial frequency.
257 Structurally related quinaxalinediones and kynurenic acid are without effect To address whether this effect of CNQX was caused by blockade of excitatory synaptic pathways, we studied the effects of NBQX (1 /~M) and DNQX (3/zM) and the broad spectrum antagonist, kynurenie acid (500 /~M). These compounds did not increase IPSC frequency (Fig. 4 and Table !). Likewise, the NMDA receptor antagonist, D-APV (50/zM), and the glycine site antagonist, 7-chlorokynurenic acid (10/zM), were also without effect on IP~.C frequency (data not shown). In' addition, the increase of IPSC frequency by CNQX persisted after pre-exposure and in the continued presence of either kynurenic acid (500 tzM) or DNQX (3 tzM). These results together suggest that the effects of CNQX on IPSC frequency are not likely to be mediated by block of conventional excitatory amino acid receptors. CNQX induces an inward current in a population of stratum radiatum interneurones Whole-cell recordings were made from interneurones located within stratum radiatum of CA3. These interneurones, which presumably make inhibitory synapses onto pyramidal cells, were identified by their anatomical location, action potential firing characteristics, and morphology, following Lucifer yellow injection. In all fourteen interneurones CNQX blocked or roduced the frequency of spontaneous EPSCs. interestingly, in six interneurones, CNQX a!s:J caused an in-
~,50]. A. Control ~, ,of ,," 2o+ ~o~r
~
o_ ol°
3 o " 4o
u>,S0 C. Recovery ==
40[ 20~t - - - -
-0"
,"o
Effect of non.NMDA receptorantagonists o n IPSCS in CA3 pyramidal nellron$
Spontaneous activity was recorded in the presence and absence of non.NMDA receptor antagonists. CNQX and DNQX were used at a concentration of 3 pM, NBQX at 1 p,M, and kynurenic acid (KYN) at 500 p,M. The statistical significance (P) and number of ceils (n) from which the data are derived are also indicated. No,a-statistically significant differences between control and drug values are represented by n.s. Values are presented as the mean ± S.E.M,
Control
Drug
P
N
Brain Research, 592 (1992) 255-260 © 1992 Elsevier Science Publishers B.V. All rights reserved ~06-8993/92/$05.00
BRES 18155
CNQX increases spontaneous inhibitory input to CA3 pyramidal neurones in neonatal rat hippocampal slices Chris J. M c B a i n , J a m e s V. E a t o n , T r a c y B r o w n * a n d R a y m o n d D i n g l e d i n e * Department of Pharmacology, University of North Carolina at Chapel Hill, Chapel Hill, NC 27599 (USA) (Accepted 19 May 1992)
Key words: CNQX; Hippocampal slice; Inhibitory postsynaptic current; Whole-cell recording; Interneuron
Whole-cell recordings were made from immature CA3 pyramidal neurones in the rat hippocampal slice. The addition of the giutamate receptor antagonist, CNQX, caused a robust increase in the frequency of spontaneous inhibitory Dost-synaptic currents (IPSC) concomitant with the expected reduction of excitatory drive to these neurones. This effect of CNQX was not shared by structurally related quinoxalinediones or ~nurcnic acid, which arc a l ~ antagonists of non-r,IMDA glutamate receptors. This effect of CNQX was abolished by tetrodotoxin suggesting that an increase in interneurone spiking was responsible for the IPSCs. Recordings from stratum radiatum interneurones of CA3 confirmed this suggestion, since some interneurones were directly depolarized by CNQX. The excitation by CNQX of a small population of stratum radiatum interneurones of CA3 complicates interpretation of experiments designed to assess the consequences of blocking excitatory transmission with this
drug.
INTRODUCTION
Certain quinoxalinediones substituted at the 6- and 7 position are competitive antagonists at non-NMDA glutamate receptors 4''~'~'14'1¢' (for review see ref. 30). The ability of the nitro-substituted quinoxalinediones, CNQX and DNQX, to block polysynaptic excitatory pathways in the rat hippocampus I'z'4'24 has made them useful tools in the study o! other receptor systems involved in synoptic transmission, particularly the NMDA and GABA receptor complexes (e.g. see ref. 1). These antagonists also dampen a variety of epileptiform tt't6-ta'23 and excitotoxic '°'2s activity in the brain. Some quinoxalinediones also block the NMDA receptor complex 3 by a competitive antagonism of the glycine co-agonist site ~'~3't6'29. Recently, Jarolimek and Misgeld t5 showed that CNQX at high concentrations (10-50 /~M) also antagonized GABA A receptors on cultured hypothalamic and mid-brain neurones. Here we describe a novel effect of CNQX not shared by other structurally related quinoxalinediones. Using the whole-cell patch clamp technique in slices of
neonatal rat hippocampi, we have observed that, at concentrations sufficient to antagonize excitatory input, CNQX causes a robust increase in frequency of spontaneous inhibitory post-synoptic currents (IPSCs) received by the CA3 pyramidal neurone population. Recordings from stratum radiatum interneurones suggest that depolarization of a small population of local interneurones is responsible for this observation. MATERIALS AND METHODS Rat hippocampal slices (300 /~m thick) were prepared from 8-15-day-old Sprague-Dawley rats, although a few cells were recorded from earlier ages, as described previously I'~. Slices were held in a submerged chamber and perfused at a rate of ~ I-3 ml/min with the following medium (in raM): NaCI 130', NaHCO.~ 24; KCI ~.5; NaH:PO4 1,25; CaCIz.2HzO 1.5; MgSO4.7tt20 1.5; glucose 10; saturated with 95% 0 : - 5 % CO: (pH 7.4, 307 mOsm) at 22-25°C. All drugs were bath applied in known concentrations by direct addition to the perfusate via a three way tap. The dead time of the perfusion system was - t rain. Tight-seal ( > I G ~ ) whole-cell recordings 7'12 were obtained from the cell bodies of 55 neurones in stratum pyramidalc of CA3, and 12 interneurones located in stratum radiatum of CA3. Patch ~lectrodes were fabricated from borosilicate glass and were not fire polished. They had resistances of 3-7 M£~ when filled with (in mM); caesium methanesulphonate 140; HEPES 10; MgCI2 2 (buffered to pH 7.3
Correspondence: C. McBain, National Institutes of Health, Bldg. 36 room 2A21, Bethesda, MD 20892, USA. *
Present address: Department of Pharmacology, Emery University School of Medicine, Atlanta, GA 30322, LISA.
256 with CsOtl, 275 mOsm). Breakthrough to the whole-cell mode was perfi+rmed under current chimp in order to provide an initial evaluation of the passive and active properties of the neurone membrane. Cells possessing an initial resting potential more negative than - 4 5 mV (pyramidal neurones - 57_+ 0.9 mV, . = 55: interneurones - 52 ± 2.5 mV. n = I I ) and overshuoting action potentials were accepted, Diabsis of the neurone by Cs + usually resulted in the cell stabilizing at a less negative potential. Neurones that fired action potentials on the anode break were discarded. Unless stated otherwise cells were voltage clamped at - 4 0 or - 5 0 mV to permit discrimination between excilato~' and inhibitory synaptic currents. Input resistances were 285 _+23 M.f~ (n = 55 pyramidal cells) and 411 ± 42 M.Q (n = I I interneurones) at this holding potential. All currents were recorded with an Axopatch-ID amplifier (Axon Instruments): re~ords were filtered at i - 3 kHz ( - 3 dB)and digitized ;it 3-10 kHz on a 80386 computer or stored on FM tape (5 kHz b:mdwidth) fi~r off-line analysis. Individual EPSCs and IPSCs were subjectively identified and accepted using the following criteria: (1) the mean current associated with the event was significantly different from that of an equivalent length of pre-cvent baseline; (2) currents were aysmmetrical with a monotonic fast rise and approximately exponential decay. Measurements u[ the decay time constant of synaptic currents r:prescnt the mean of at least 15 individual events from each aeurone and were fit by one exponential component (simplex algorithm, least squares criteria), Amplitude histograms were constructed from individually measured IPSC amplitudes and hinned in 2 pA intervals. The series resistance was determined by balancing the "bridge' in current chimp, and lypically ranged from 10-3() M.f2, The amplitude of the inhibitory synaptic currents was never in excess of 80 pA, thus a series resisl.'lnce of 25 M.f~ would rcsuh in a series resistance error of no greater thai} 2 mV. Membrane potentials were not corrected for these errors, The steady-state current-voltage (I-V) relation of hath-applied CNQX was obtained by measuring the current ;It the end of 25()-I,1}()() ms voltage st~ps applied in = 10 mV increments at 10 s inlcrvals, A holding polential o f ~1=5|l n'lV W,'IS selected for Hlese stltdh.~s to inacliV;|le most voltagedependent currents. I=V rehithms were obt;fined prior to, durin~ :tddilion and fi)llowing rentoval of k;finate, The mean of Ihc control and recovery was Ihun sul'qra¢led from that in CNQX, All data nr¢ represented ,'is the: mean i S.E,M, Where approprt. ate. a Stndcnt's t.test wits pcrfornlcd, I)rugs used wcrc tetrodotoxin ('rl'X, Si~,nla), hicttcuIline nl~tho. I','omide (Sigma }. h.cyarlo.7.nitro.quinoxalinc.2,3.dione (('NQX} and h.7.dinit|oquitloxuline.2,3.dione (DNQX, Toeris Neuramin), kynurenic a~.'id, (Signla), and 2,.~.dihydroxy.6-nilro.7.sulfin'noyl. benzo(F) quinoxalinc (NBQX, generous gift from T, I lonore):
RESULTS CN(.)X increases spontaneous IPSC fi'equency hz CA3 pyramids Pyramidal ncuroncs were voltage clamped at - 40 to - 5 0 mV, which permitted the separation of inhibitory (outward at this holding potential) from excitatory (inward) synaptic currents (Fig, 1). The pyramidal ncutones received spontaneous IPSCs at a frequency of 0.3-17.1 Hz (mean 3.4:t:0,6 Hz, . = 3 4 ; Table I). These IPSCs ranged in amplitude from 5 to 80 pA. The falling phase o1" the IPSC could be fit by a single exponential with a time constant of ~ 20 ms (Table !), EPSCs were not studied further and have been described in detail eisewheret". In 3(} of 34 neurones addition of CNQX (3 pM) to the bathing medium virtually abolished EPSC activity while increasing the frequency of IPSC activity (Figs. I
>,70t
A. Control
60f 5of
.
. .
=o',+of o 20f
cn 1Of
E.itlii
Q- | ~~ __ 00 10 20 30 40 50 - -
:=,.,70 B. 3.aM CNQX
sot g'+ot 3of
u2°t
~nlo+
0
I l L ..lit.
~
_
10 20 3'3 40 50
"
E+'O~lS°sI°l
~70,C. 3#M CNOX+ I/zM Bicuculline
,,.-:3ot o 2o1
u~ 10,1.
OOL 10 20 3 0 40 IPSC Amplitude (pA)
50
Fig, 1, CNQX increases the frequency of spontaneous IPSCs and blocks spor{aneous EPSCs. A representative recording from a stratum pyramidale CA3b neurone. IPSC amplitude histograms (bin size = 2 p a l are shown on the left. The histograms were generated from 59 s of data. In the current traces shown on the right, upward deflections represent IPSC:s and downward deflections represent EPSCs, Those evenls marked ' + ' were accepted as IPSCs, CNQX induced :k suhst:mlial increase in IPSC frequency while having no noticeable effect on IPSC amplitude or decay time constant, A: control (ACSF only); m~an amplitude ,= 17 pA, ntean decay constant ., 23 ms, IPSC frequency + 3,8 |lz, B: 3 / ~ M CNQX; IPSC amplitude ,~ 21 pA, decay lint~ constant +. 24 ms, IPSC frequency =9.2 llz, C: most IPSCs arc clitniflated by the addition of bicuculline (5 # M)in the presence of C'NQX,
an0 2). This effect of CNQX was not accompanied by a~y change in the pyramidal neurone holding current or input resistance, and CNQX was without significant effect on the r~lean amplitude or decay time constant of IPSCs (Table !). This effect was readily reversible and could be reproduced by re-application of CNQX (Fig. 2). The synaptic currents in CNQX could be blocked by the addition of the specific GABA^ receptor antagonist, bicuculline (5/zM; Fig. IC), or tetrodotoxin (Fig. 2El, suggesting that the IPSCs were caused by ongoing interncurone activity. The magnitude of the increase in IPSC frequency by CNQX varied greatly from cell to cell, from less than 2-fold to about 25-fold. A plot of IPSC frequency in the absence of CNQX vs. that in the presence for all cells tested (Fig. 3) indicates that this wide range observed is not due to a 'ceiling effect'. Rather, CNQX appeared to increase IPSC frequency by 2 - 3 / s regardless of the initial frequency.
257 Structurally related quinaxalinediones and kynurenic acid are without effect To address whether this effect of CNQX was caused by blockade of excitatory synaptic pathways, we studied the effects of NBQX (1 /~M) and DNQX (3/zM) and the broad spectrum antagonist, kynurenie acid (500 /~M). These compounds did not increase IPSC frequency (Fig. 4 and Table !). Likewise, the NMDA receptor antagonist, D-APV (50/zM), and the glycine site antagonist, 7-chlorokynurenic acid (10/zM), were also without effect on IP~.C frequency (data not shown). In' addition, the increase of IPSC frequency by CNQX persisted after pre-exposure and in the continued presence of either kynurenic acid (500 tzM) or DNQX (3 tzM). These results together suggest that the effects of CNQX on IPSC frequency are not likely to be mediated by block of conventional excitatory amino acid receptors. CNQX induces an inward current in a population of stratum radiatum interneurones Whole-cell recordings were made from interneurones located within stratum radiatum of CA3. These interneurones, which presumably make inhibitory synapses onto pyramidal cells, were identified by their anatomical location, action potential firing characteristics, and morphology, following Lucifer yellow injection. In all fourteen interneurones CNQX blocked or roduced the frequency of spontaneous EPSCs. interestingly, in six interneurones, CNQX a!s:J caused an in-
~,50]. A. Control ~, ,of ,," 2o+ ~o~r
~
o_ ol°
3 o " 4o
u>,S0 C. Recovery ==
40[ 20~t - - - -
-0"
,"o
Effect of non.NMDA receptorantagonists o n IPSCS in CA3 pyramidal nellron$
Spontaneous activity was recorded in the presence and absence of non.NMDA receptor antagonists. CNQX and DNQX were used at a concentration of 3 pM, NBQX at 1 p,M, and kynurenic acid (KYN) at 500 p,M. The statistical significance (P) and number of ceils (n) from which the data are derived are also indicated. No,a-statistically significant differences between control and drug values are represented by n.s. Values are presented as the mean ± S.E.M,
Control
Drug
P
N
